Tambo AI
An open-source React toolkit for building agents that render generative UI: components are registered with schemas, and an agent chooses and streams the right one as it responds to a request.
Tambo AI is an open-source toolkit for building agents that render user interfaces, known as generative UI, in React applications. A developer registers existing React components together with Zod schemas; the agent then picks the right component for a request and streams its props to it as the model generates them, so an interface updates live instead of only returning text.
Tambo includes an LLM conversation loop, so a project supplies its own API key for OpenAI, Anthropic, Gemini, Mistral or any OpenAI-compatible provider, and can optionally use agent frameworks such as LangChain or Mastra on top. Streaming, cancellation, error recovery and reconnection for the props stream are handled for the developer. The backend that manages conversation state and agent execution can run as a hosted cloud service or be self-hosted with Docker on a project's own infrastructure.
It ships with a pre-built component library of generative UI primitives and starter templates, including a chat interface and an analytics dashboard, and is written in TypeScript under the MIT license.